GitLab

Depending on StringProperty's display type, its property widget shows either just the QLineEdit, or additionally a browse button that pops up a QFileDialog for either loading or saving a file, or choosing a directory.

closes #43

Refactor conversion api

refs #474
refs #553

* Introducing ImageRepresentationTest
* Improved ImageDataTest
* Silenced Debug output in test-campvis

Merging GoogleTest

Issue 546

closes #546
closes #37

Fixed and completed tgt::Texture::texelAsFloat() methods to support all data types and number of channels.

refs #37

* Implemented trackball navigation interaction for rendered geometries
* Implemented correct rendering of textures with correct aspect ration (no longer weirdly stretched textures)
* removed obsolete (because redundant) renderFullscreen option

refs #546
refs #37

* DataContainerInspectorCanvas uses the GeometryRenderer processor to render geometry
* DataContainerInspectorCanvas::_textures now uses DataHandle to improve thread safety avoid segfaults
* replaced some function calls with Qt signals to loose coupling
* removed obsolete fields and methods

refs #546

refs #474

Statically linked builds strip all unused objects from their sources. This would be the case for the converters since they are nowhere called explicitly. In order to fix that, we use the same pattern as with the pipeline registration: The CMake build scripts parse all headers for an explicit template instantiation of a ConversionFunctionRegistrar. All these headers are included from the generated gen_converterregistration.h file, which is itself included from imagerepresentationconverter.cpp and thus gets compiled into the final executable also on static linking.

refs #474
refs #553

get rid of gcc warning from test-campvis. however, there is one remaining in gtest-internal.h, and a few in concurrenthistogram.h. couldn't figure out the way around for those

Fixed GenericGeometryTransferFunction not setting the clear color correctly before rendering into TF texture.
This also fixes the broken IXPV pipeline (Issue #62) that got also cleaned with this commit.

closes #62

* Reverted some of the changes introduced with 3c7bf8c0: tgt::FramebufferObject now uses EXT functions (instead of core functions) throughout the entire code. Previous mix of EXT and core function calls showed problems with AMD GPUs. All core methods broke results on Intel GPUs.

google test merge request feedback reviewed. addressed all necessary major and minor issues that were pointed out; and hopefully fixed my stupidity in LinearMapping::operator==

Removed all code fragments in core code that was ITK specific (conversion from/to GenericImageRepresentationItk). The functionality was replaced by the new conversion functors in imagerepresentationconversionitk.h and imagerepresentationconversionitk.cpp.

refs #553
refs #474

Conversions between image representations are now managed at one central place: The ImageRepresentationConverter singleton uses the proven and established registration through static template instantiation idiom to register conversion functors during static initialization. Therefore, the ConversionFunctionRegistrar registers a conversion functor to a target representation type.

As proof-of-concept implementation, the former conversion API through T::tryConvertFrom, where T is a specific image representation, has been converted to the new API and merged into imagerepresentationconversioncore.h providing a conversion functor for each campvis-core representation.

Furthermore, implemented conversion from ImageRepresentationGL to GenericImageRepresentationLocal<>.

refs #553
refs #474

rewritten init() with the glcontextmanager instead of qtcontextmanager, updated ImageData, NumericProperty, GenericProperty test codes

added the documentation (brief). added test class for StringUtils. trimmed out the redundant code segments.

added DataContainer, StringProperty aka GenericProperty, NumericProperty test code and reorganized the folders

"MERGED 4 commits - 1. cleaned up googletest 2. solved std_tuple issue with four modified files from current beta googletest 3. MTd-MdT issue resolved with CMake flag 4. googletest produces two projects named gtest and gtest-main. a third project is created linked with gtest-main and all our testing libraries"

* Geometry1DTransferFunctionEditor not correctly drawing image intensity histogram.
* VolumeExplorer not always initializing slice- and raycasting render size properties.

Lightsourcedata

Removed ProcessorDecoratorShading and replaced all of its usages with the new LightSourceData API. This also includes adding LightSourceProviders to the corresponding pipelines.

refs #547

* Introducing base module with LightSourceProvider processor
* Updated GeometryRenderer and GeometryRendererDemo to use LightSourceData instead of ProcessorDecoratorShading
* Introducing LightSourceData::bind()

refs #547